• This short tutorial can't cover the vast R universe
• But… R has vibrant community supporting it!
• You can find endless tutorials, books, presentations, etc. on just about any aspect of R you want – from beginner topics to advanced matters.
• Goal for today:

• RStudio provides many online learning resources
• And even more helpful links buried in an RStudio wiki page
• Code School has an excellent tutorial to help you get acquainted with R
• StackOverflow is a great place to find answers to R programming questions
• Google Developers intro to R video tutorials
• Advanced R by Hadley Wickham
  • The first few sections are excellent for beginners
  • The rest of the book is essential for a deep understanding of R

Here's some useful tips on using the R console in RStudio.

• The R console can be used as a calculator
• Enter the following commands into your R console to see it in action:

2 + 2
5 * 4
2^3

• It is also smart enough to understand order of operations:

2 + 3 * 4/(5 + 3) * 15/2^2 + 3 * 4^2

> y
Error: object 'y' not found

You can name your variables anything you want, but there are a few rules:

1. Names may contain letters, numbers or periods, and the underscore character
2. Names must start with a letter or period.
3. v.one and v_one are valid names but v one is not (because it includes a space).

More information on general R programming style can be found here.

1. Vectors
2. Matricies/Arrays
3. Data Frames
4. Lists

Vectors are R's most basic data structure. When we created the variable x in the previous section, we had actually created a vector of length 1. The elements contained in a vector must be of the same type (see prev. section). Vectors including more than one element are frequently constructed using the c() (concatenate) function:

a <- c(1, 2, 5.3, 6, -2, 4)  # numeric vector
b <- c("one", "two", "three")  # character vector
c <- c(TRUE, TRUE, TRUE, FALSE, TRUE, FALSE)  # logical vector

You can access values inside of a vector by subsetting it with the [] operators. Examples:

> a[1] # access the first element of a
[1] 1
> a[3] # access the fourth element of a
[1] 5.3
> b[length(b)] # acess the last element of b
[1] "three"

Matrices and arrays are vectors with dimensions. Since they are vectors, they can only contain elements of the same type. Matrices have 2 dimensions (rows & columns) and are created with the matrix function:

> matrix(1:25, ncol = 5, nrow = 5)
     [,1] [,2] [,3] [,4] [,5]
[1,]    1    6   11   16   21
[2,]    2    7   12   17   22
[3,]    3    8   13   18   23
[4,]    4    9   14   19   24
[5,]    5   10   15   20   25

We can subset arrays in matrices in the same way we did vectors, but we must be conscious of their dimensionality. For a matrix, m, we would access the data stored in row 2 and column 5 by typing m[2, 5].

• Data frames are the R data structure you will deal with most frequently
• R's data import/export functions will return these by default
• They can also be created using the data.frame() function
• Here's a quick example creating some dummy experimental data:

> data.frame(group = rep(c('trt', 'ctrl'), each=4), 
            response = rnorm(8))
   group    response
1    trt  0.01231700
2    trt  0.55028627
3    trt -0.24411263
4    trt  0.35996580
5    trt  0.89960231
6   ctrl -0.76686861
7   ctrl  0.03573818
8   ctrl  1.69159332

• Lists are containers allowing to store multiple R objects in the same parent object
• Unlike vectors, enclosed data can have different types
• As an example, we'll create a list with some info about a person:

> person <- list(first = 'andrew', last = 'borgman', age = 25)
> person
$first
[1] "andrew"

$last
[1] "borgman"

$age
[1] 25

• R provides users with many built in functions
  • These allow users to perform a range of tasks
  • Computing summary statistics, to fitting linear models, to generating random variables.
• There are lots of good lists of functions out there
• In most cases, R functions have been given logical names.
• Ex: the mean function is used to find the average of a vector of numbers:

> my_data <- c(10, 22, 44, 55, 14, 66)
> mean(my_data)
[1] 35.16667

• Pre-RStudio, all of this searching took place on the command line
• You may find this is a more convenient way to search for R help
• The help() and ? functions can be used to search the documentation for functions matching the inputted text
• The following functions both search R's help documentation for any functions named mean

?mean
help(mean)

• But sometimes we don't know the name of the function we seek
• Fortunately, R's documentation search is flexible enough to accommodate that
• Let's try to find a function that lets us compute the standard deviation of a vector of numbers.
• First, we'll try a search with the name of the function used in Excel software

> ?stdev
No documentation for ‘STDEV’ in specified packages and libraries:
you could try ‘??STDEV’

• You'll see that R gives us a hint and is telling us to use the function ??.
• This function works like the search box we used to find help in RStudio. Let's search for "standard deviation":

??"standard deviation"

Entering this, we see the same kind of search results returned in the RStudio help pane.

And, if all else fails, try Googleing to see if the function you seek exists!!

The R workspace can be thought of as a container holding all of the objects you've created duing your R session. You can print a list of all of the objects in your current workspace using the ls() function. If we start a new R session, our workspace will be empty:

> ls()
character(0)

And we'll be able to see some objects if we add them:

> x <- 20
> y <- 30
> z <- x + y
> df <- data.frame(nums = 1:10, b = letters[1:10])
> ls()
[1] "df" "x"  "y"  "z" 

• save.image: Lets you save a snapshot of your entire workspace to a *.RData file.
  • Typical usage: save.image('my-work.RData')
• save: Lets you save a snapshot of a few specified objects to a *.RData file.
  • This can be useful if you only want to save one or 2 things (like a modified data set).
  • Typical usage (saving df the data.frame we created in the last section): save(df, file = 'my-dataframe.RData')
• load: Lets you load your saved *.RData files back in to R ot continue working on them.
  • Typical usage (saving df the data.frame we created in the last section): load('my-dataframe.RData')

• There are two main ways to clear your workspace in R/RStudio
  1. using the rm() function to remove objects and
  2. using the Clear button in RStudio.
• The rm() command allows you to selectively remove objects from your R session when they are no longer needed
• Example of the usage:

> ls() # start in an empty workspace
character(0)
> y <- 1
> z <- 1
> ls() # can see the two objects we created
[1] "y" "z"
> rm(y) # remove y
> ls()  
[1] "z"

You can remove all objects in your workspace by using ls() to generate a vector of all the objects that have been created, and passing that to the rm() function:

rm(list = ls())

You can also use a button in RStudio's Environment panel to remove all of the objects in your workspace. RStudio will prompt you asking if you are sure you want to go through with deleting all objects, choosing Yes will permanently delete all objects in the workspace.

• Packages are what make R so great.
• Packages allow R users to develop and share new code to extend the language
  • e.g. implementing a new statistical model for analyzing microarrays, creating an awesome new graph to show copy number alterations in tumor samples, etc.
• R has become the lingua franca for statisticians developing new statistical techniques
• Has also been widely adopted in the bioinformatics community (Bioconductor)

Most generic R packages are hosted on the Comprehensive R Archive Network CRAN. To install one of these packages, you would use install.packages("packagename"). You only need to install a package once, then load it each time using library(packagename). Here's how one would install and load the ggplot2 package.

# Install only once.
install.packages("ggplot2")

# Load the package every time you want to use it.
library(ggplot2)

Now that we're feeling a bit more comfortable with the R environment, we'll explore how we can import our own experimental data into R.

Before going any further, please download the zip file below and extract its contents somewhere on your computer.

• data_dl.zip

The easiest way to store data for import into R is in text files. R has facilities for importing data directly from Excel spreadsheets or from some data base format, but such performing such operations is outside the scope of this course.

Flat text files containing data typically have some type of special character (delimiter) – think: tab, semicolon, space – for separating different columns of data. Data stored in text files most commonly falls into the following formats.

• Tab-delimited files are easily exported from Excel
• One of the best ways to store data in a flat text file
• Typically have a file extension of .txt or .tsv.

ID  Group   Gene1   Gene2   Gene3   Gene4
Sample 1    Group 1 9.695   4.694   3.733   4.874
Sample 2    Group 1 8.087   3.276   2.220   3.095
Sample 3    Group 1 9.885   6.297   5.842   8.233
Sample 4    Group 1 7.832   -1.286  2.594   -1.089
Sample 5    Group 1 10.239  4.474   3.300   3.377

File -> Save As -> Tab-Delimited Text

• It is also very easy to export comma-delimited files from Excel
• Comma-delimited files typically have a file extension of .csv.

"ID","Group","Gene1","Gene2","Gene3","Gene4"
"Sample 1","Group 1",9.695,4.694,3.733,4.874
"Sample 2","Group 1",8.087,3.276,2.220,3.095
"Sample 3","Group 1",9.885,6.297,5.842,8.233
"Sample 4","Group 1",7.832,-1.286,2.594,-1.089
"Sample 5","Group 1",10.239,4.474,3.300,3.377

File -> Save As -> Comma Separated Values

• Just saw RStudio generate a read.delim function to import our .tsv data frame
• read.delim is a special alternative to R's more general read.table function (?read.table for details)
• read.csv is another special alternative to read.table
• read.csv & read.delim have

• Oftentimes, analysts will want to work with only a subset of their data at one time
• A definitive guide on subsetting in R can be found here
• There are many ways to subset data in R, but we'll touch on a couple important ones

> gene.exprs.long[1:5, ] # print the first 5 rows
        ID   Group  Gene Expression
1 Sample 1 Group 1 Gene1   9.695228
2 Sample 2 Group 1 Gene1   8.087463
3 Sample 3 Group 1 Gene1   9.885696
4 Sample 4 Group 1 Gene1   7.832890
5 Sample 5 Group 1 Gene1  10.239599

> gene.exprs.long[1:5, 1:2] # print the first 5 rows and columns 1 & 2
        ID   Group
1 Sample 1 Group 1
2 Sample 2 Group 1
3 Sample 3 Group 1
4 Sample 4 Group 1
5 Sample 5 Group 1

• Usage: subset(dataset, logical conditions)

> subset(gene.exprs.long, ID == 'Sample 1') # All Sample 1 measures
          ID   Group  Gene Expression
1   Sample 1 Group 1 Gene1   9.695228
51  Sample 1 Group 1 Gene2   4.694323
101 Sample 1 Group 1 Gene3   3.733354
151 Sample 1 Group 1 Gene4   4.874305

> subset(gene.exprs.long, Expression > 10) # gene expression > 10
          ID   Group  Gene Expression
5   Sample 5 Group 1 Gene1   10.23960
6   Sample 6 Group 1 Gene1   11.01402
10 Sample 10 Group 1 Gene1   10.10460
13 Sample 13 Group 2 Gene1   10.21796
16 Sample 16 Group 2 Gene1   10.03342
46 Sample 46 Group 5 Gene1   10.11114

• Data analysts have come up with ways to describe the format of the data they work in
• Most of the data you come into contact with will be in either wide or long format
• See data files in example zip folder for wide and long versions of same data set

There are three main graphing frameworks available in R for creating high quality plots:

1. Base Graphics
2. Lattice Graphics
3. ggplot2 Graphics

I would suggest learning ggplot2

ggplot2 is a plotting system for R, based on the grammar of graphics, which tries to take the good parts of base and lattice graphics and none of the bad parts. It takes care of many of the fiddly details that make plotting a hassle (like drawing legends) as well as providing a powerful model of graphics that makes it easy to produce complex multi-layered graphics (taken from ggplot2 site). ggplot2 allows intro R users to create high quality data visualizations with little effort using an intuitive plotting syntax.

• Nice walkthrough of using qplot to make complex graphics
• Excellent quick overview of ggplot2
• In depth guide of ggplot2
• Quick reference for ggplot2